The Impact of Credit Risk on Profitability of the …

urnal of B

Jo

Affairs

cial

an

ss & Fi

n

ine

us

ISSN: 2167-0234

Journal of

Business & Financial Affairs

Saeed and Zahid, J Bus Fin Aff 2016, 5:2

DOI: 10.4172/2167-0234.1000192

Research Article

Open Access

The Impact of Credit Risk on Profitability of the Commercial Banks

Saeed MS* and Zahid N

Glasgow Caledonian University, Glasgow Business School, Glasgow, UK

Abstract

This paper aimed to analyse the impact of credit risk on profitability of five big UK commercial banks. For measuring

profitability, two dependent variables ROA and ROE were considered whereas two variables for credit risks were: net

charge off (or impairments), and nonperforming loans. Multiple statistical analyses were conducted on bank data from

2007 to 2015 to cover the period of financial crisis. It was found that credit risk indicators had a positive association with

profitability of the banks. This means that even after the deep effects of credit crisis in 2008, the banks in the UK are

taking credit risks, and getting benefits from interest rates, fee, and commissions etc. The results also reveal that the

bank size, leverage, and growth were also positively interlinked with each other, and the banks achieved profitability

after the financial crisis and learned how to tackle the credit risk over the years.

Keywords: Credit risk; UK commercial banks; Bank profitability;

Net charge off; Nonperforming loans; ROA; ROE

Introduction

The banking industry worldwide has been more complex over the

years because of rapid development and growth of financial security

market [1]. Consequently, the banks started to practice multiple

compound operations without even perceiving the risks associated

with these transactions. As a result, the risk attitude and risk exposure

of banks became more composite and subject to system failure and thus

they caused to break down to economic system of the country where

they operate. The governments of various countries tried to control

the situations by practicing regulatory reforms in order to stabilise the

economy. However, it is worth to declare that these reforms did not

work well and ended up with the similar outcomes such as financial

volatility and economic downturn around the globe including UK,

USA and other economies on which worlds¡¯ institutions are based

to great extent. During all the circumstances the most exposed risk,

which was difficult to discover, was the credit risk [2]. The significance

of credit risk is enlarged by the reality that it is associated with the

collateral problem. Hence, it becomes the most controversial topic to

be discussed and explored. For dealing with credit risk, Basel II also

practiced and adopted different credit risk management techniques

[3]. The primary objective of these practices was to improve the quality

of credit risk management without limiting the competitiveness of the

banks worldwide.

Over the last 10 years, the quality of the loan and its portfolios

across many economies worldwide stayed comparatively stable until

the emergence of 2007-08 financial crises. Since then the quality of the

bank assets declined quickly because of the world economic downturn.

The reality is that the loan performance is closely associated with the

economy of any country and decline in the loan performance was not yet

standardised across the world economies [4]. For instance, some crosscountry analyses and evaluations in terms of GDP performance during

the time of crises reveal extensive enlargements in non-performance

loans. In 2009, the economy of Latvia squeezed by 18 percent decline in

GDP. Simultaneously, the economy of Germany also shrank by nearly

5% in terms of GDP and when it appears to non-performing loan ratio,

it also contracted by great extent [5].

Banks face too many serious problems due to unsuccessful credit

risk management but the credit lending remains the chief activity of

the banking sector throughout the world. The core cause behind it that

banks can no longer survive without this activity. This is the reason

J Bus Fin Aff

ISSN: 2167-0234 BSFA an open access journal

that credit worth is considered as a key sign of financial health and

soundness of financial institutions particularly the banks. The interests

charged by the banks on advances and loans shape large part of the

bank¡¯s assets and delays and defaults of credits and advances create

solemn circumstances for both the lenders and borrowers and even

the whole economy can be disturbed as evident in the 2008 financial

crisis. Different studies in the context of banking crises across the

world uncover the fact that poor credits (asset quality) are the primary

cause of failure of the banks [6]. Stuart indicates that the ratio of nonperforming loans (bad loans) all around the world was extremely high

between 1999 and 2009 in commercial banking sector [7]. And this

was due to a number of reasons such as absence or inadequate loan

collaterals, poor loan processing, ineffective credit risks management,

excessive intervention during loan lending procedure, and several

negative impacts on bank profitability.

Therefore, by considering the importance of credits in the banking

sector and their severe economic impact, it is extremely important

to find the relation and impact of credit with/on profitability of the

bank. The banking theory points out 6 major risks associated with the

credit policy of banks. These risks are: credit risk (or repayment risk),

credit deficiency risk, operating risk, portfolio risk, interest risk, and

trade union risk [8]. However, credit risk is the most vital risk among

them and thus, it requires special awareness and concentration. Hence,

a sincere attempt is made in this dissertation to make the modest

contribution to the credit risk literature by analysing the impact on UK

banking sector with particular focus on five big UK commercial banks

including HSBC, Barclays, Royal Bank of Scotland, Lloyds Banking

Group, and Standards Chartered Bank.

Literature Review

Figure 1 illustrates the proposed theoretical model of this study.

The model consists of two profitability indicators (ROA and ROE)

*Corresponding author: Muhammad Sajid Saeed, Glasgow Caledonian

University, Glasgow Business School, Glasgow, Cowcaddens Rd, Lanarkshire G4

0BA, Scotland, UK, Tel: +441413313000; E-mail: msaeed14@caledonian.ac.uk

Received June 17, 2016; Accepted June 27, 2016; Published June 30, 2016

Citation: Saeed MS, Zahid N (2016) The Impact of Credit Risk on Profitability

of the Commercial Banks. J Bus Fin Aff 5: 192. doi:10.4172/2167-0234.1000192

Copyright: ? 2016 Saeed MS, et al. This is an open-access article distributed

under the terms of the Creative Commons Attribution License, which permits

unrestricted use, distribution, and reproduction in any medium, provided the

original author and source are credited.

Volume 5 ? Issue 2 ? 1000192

Citation: Saeed MS, Zahid N (2016) The Impact of Credit Risk on Profitability of the Commercial Banks. J Bus Fin Aff 5: 192. doi:10.4172/21670234.1000192

Page 2 of 7

Bank size

Profitability indicators

Growth

Net charge off

over total loans

ROA & ROE

Nonperforming

loans over total

loans

Leverage

Figure 1: Theoretical model.

which are considered as dependent variables; and five independent

variables (bank size, growth, leverage, and credit risks (ratio of net

charge off and non-performing loans to total loan) (Figure 1).

There are several risks linked with the banking sector namely credit

risk, earning risk, interest rate risk, market risk and liquidity risks

are key risks. There are three dominant categories of these risks like,

credit risk, operations risk and market risks. Among all types of risks

the vigorous part is played by the credit risk without any suspicion

that bank¡¯s biggest asset is loan which is normally consists of 50 to

70 per cent of banks value. Credit risk is well-defined by the Basel

committee on Banking supervision as ¡°potential that a bank borrower

or counterparty will fail to meet its obligations in accordance with

agreed terms¡± [9].

The credit risk based upon the obtainable internal data which is

measured by investigating the adjustments in quality loans (medium

or low) over total asset ratio. The credit risk can be controlled and

dropped down by the chance provided by this ratio. However several

scholars have stated the conventional ratios which can be employed

to recognise the credit risk if no data about medium loan quality is

available, for example;

o

Total loans to total deposits

o

Total loans to total assets

o

Nonperforming loans to the total loans

o

Nonperforming assets to total loans and advances

o

Loan loss reserves to the total loans

o

Net charge-offs of loans to total loans and advances [5,10-12].

Literature on non-performing loans has extended along with the

attention towards investigating the major reasons behind the financial

vulnerability over the last few years. Financial weakness is primarily

because of critical role impaired assets have, proven by the evidence

which shows the firm link between banking/financial crises and NPLs

in Sub-Saharan African countries and East Asian countries during

the 1990s. The prevailing literature to scrutinise the determinants of

non-performing loans in Guyana is studied in the current section to

formulate a theoretical framework.

Keeton and Morris made an examination on the causes of loan

losses in their earliest study [13]. Latterly they estimated the 2470 losses

insured by commercial banks in the US during the time period of 1975

to 1985. NPLs are used by them as the prime method of calculating loan

gains and losses. The findings of their study indicate the variation in loan

losses documented by banks is mainly described by the local economic

situations and inadequate performance of particular industries.

The publication of Keeton and Morris is followed by several other

studies which anticipated interrelated reasons for problems regarding

credits in the United States. Another study was conducted by Sinkey

J Bus Fin Aff

ISSN: 2167-0234 BSFA an open access journal

and Greenwalt on the loan gain/loss experience of major banks in the

United States [14]. The results of their study postulate that Loan ¨C loss

rate is affected by both internal and external factors of these banks.

The term loan loss rate refers to an addition of NPLs and net loan

charge offs and dividing this sum by net charge offs plus total loans.

The main findings of their study showing that internal factors namely

High rate of interest, Excessive lending and volatile funds significantly

and positively influence the loan-loss rate. Sinkey and Greenwalt

also discovered that loss rate of banks is also based on the economic

conditions. They used the data of giant banks in the US during the

Time period of 1984 to 1987 by employing simple log-linear regression.

Another study on the effect of loan expansion in the United States

is conducted by Keeton [13]. The author took the data from 1982

to 1996 by employing the regression model for empirical analysis.

Evidence show that credit growth is strongly associated by impaired

assets stated by the author. The major credit loss in particular US States

is affected by the rapid credit growth that has a relationship with lower

credit standards.

Many studies also provided similar results conducted other than

US financial system. Kodan and Chhikara examined the Indian

banking industry through statistical tools and techniques by analysing

the trends and composition of NPAs [15]. The analysis of data showed

the significant reduction in NPAs in Indian industry over the time.

Salas and Saurina conducted a study on the Spanish commercial

and savings banks by using a comprehensive dataset and framework

for 1985 to 1997 [16]. The key aim of this research was to explore

the determinants of problem credits in Spanish banks. Their findings

indicate that major variation in NPLs is mainly explained by market

power, capital ratio, bank size, rapid credit expansion, and true

GDP growth. Similarly in another study Spanish banking sector was

investigated for the period of 1984 to 2003. The empirical evidence

shows that loan terms, higher interest rates, and GDP growth are

key determinants of NPLs. This study points out that managers in

commercial banks provide more loans when economic conditions

are excellent and trigger several issues such as agency problems, herd

behaviour and disaster myopia.

Rajan and Dhal conducted a study on the commercial banks of

India by utilising panel regression analysis [17]. They reported that

encouraging economic situations and other financial indicators such as

credit direction, bank size, credit terms, and maturity have significant

influence on the NPLs of Indian commercial banks.

Fofack made an attempt to examine the determinants of NPLs

in several countries of Sub-Saharan Africa by using a pseudo panelbased model [18]. The researcher found that NPLs are determined by

many factors such as interest margins, interest rate, economic growth,

interbank loans, and exchange rates etc. The NPLs and these economic

factors have important role to play in the undiversified African

countries.

The study of Hu et al. indicates the association between commercial

bank¡¯s ownership structure and NPLs in Taiwan for the period of

1996 to 1999 using panel dataset [19]. The results indicate that Nonperforming loans are negatively associated with higher government

ownership and bank size while diversification seems indifferent and

may not be a determinant.

Ratio analysis is used to measure and analyse the bank¡¯s profitability.

Financial statements of banks demonstrate some ratios and some can

be calculated based on requirements if needed. Koch and MacDonald

Volume 5 ? Issue 2 ? 1000192

Citation: Saeed MS, Zahid N (2016) The Impact of Credit Risk on Profitability of the Commercial Banks. J Bus Fin Aff 5: 192. doi:10.4172/21670234.1000192

Page 3 of 7

stated that relatively appropriate measures for measuring the bank¡¯s

profitability level are Return on Assets (ROA) and Return on Equity

(ROE) [20]. These measures are described in the light of the existing

literature in this section.

ROA is calculated as a percentage of net income and total assets.

ROA is used as main profitability measure in most of the organisations

including banks and financial institutions. The ROA demonstrates the

level of net income produced by the bank and also determines how

the assets utilised by banks to generate profit over the years [6]. The

competence and proficiency of banks in transforming their assets into

profits is also indicated by it. Hence, to improve the performance of

banks, they always attempt to achieve higher ROA. The ranking of

banks is usually based upon the higher ROA ratio and total assets. As

a general view, particularly in banking sector, ROA is known as good

profitability multiplier for the reason that equity multiplier does not

influence it [21].

A percentage of net income over shareholder¡¯s equity is termed as

ROE. The net income comprised of all types of earnings like preferred

stock income, surpluses, undivided profits and capital reserves. The

difference between net assets and liabilities is termed as shareholder¡¯s

equity on the other hand. The most common measure to determine the

effectiveness of banks of generating revenue based on every element of

shareholder¡¯s equity.

To attain sufficient level of profitability, Both ROE and ROA refer

to bank¡¯s managerial ability. According to Golin and Delhaise, the

ROE between 15 to 20 per cent is considered to be good for a banking

institution [6].

The significant difference between ROA and ROE measures is

debt. The total assets and shareholder¡¯s equity will become equal in the

absence of debt; consequently the results drawn from each measure

would be equivalent. According to the Koch and MacDonald, a greater

value of ROE is not always considered as inspirational indicator of

good performance of the bank, consequently ROA is known as suitable

measure of profitability and efficiency of the banks [20].

An extensive stock of earlier literature has discussed the ROE as a

significant indicator to quantify the profitability of the banks. Foong

revealed that ROE is used to measure the efficiency of banks which

explains to make upcoming profits; the reinvested income is used to

what extent [22].

According to Riks bank¡¯s Financial Report to define the profitability

in the banks, the technique which is normally used is to associate profit

with shareholder¡¯s equity [23]. Moreover, in the paper ¡°Why Return on

Equity is a Useful Criterion for Equity Selection¡±, the author has found

a very useful instrument for profit generating efficiency provided by

ROE for its ability to measure the extent of company¡¯s earnings on the

equity capital.

Company¡¯s after tax annual net income divided by shareholder¡¯s

equity is termed as ROE. NI is the deduction of all expenditures and

taxes from total earnings. Retained earnings added to capital invested

in the company are called equity. Basically, the amount of earnings

made from equity is termed as ROE. The higher value of ROE indicates

that, without injecting new capital into the company profit is rising.

Each year shareholders are provided with more of their investment

referred by a progressively growing ROE. Conclusively, the greater

ROE is fruitful for them as well as for the growth of the company.

Additionally, ROE guides the investors how efficiently the capital is

reinvested by taking the retained earnings.

J Bus Fin Aff

ISSN: 2167-0234 BSFA an open access journal

According to the study of Waymond, the indicators widely used

with greater esteem for credit analysis in banks is profitability ratios, as

results of management performance is associated with the profitability

[24]. Most widely used ratios are ROE and ROA, and the ROE level of

quality ranges from 15-30 percent and at least 1 per cent for ROA.

Joetta suggested that the aim of ROE as the investigation of total

profit produced by the firm¡¯s equity. It is also mentioned that to

engender profit from equity the ROE is used as a gauge of the efficiency

[25]. This ability is associated with how accurately the collaterals are

utilised to yield the earnings. The assets¡¯ quantity produced by the

company against each equity dollar, considerably determined through

the effectiveness of assets utilisation. Thus, after bringing the evidence

of ROE used as the profitability indicator, the discussion can be moved

towards credit risk management indicators.

Research Methodology

Research design

The research design embraces the methods on which the research

work is founded on Saunders et al. [26]. In other words, it can be said

that it is composed of the type of the study which is employed by the

researchers to accomplish the objectives. The type of study covers

various aspects such as hypotheses, variables, methods, and analyse

framework. Descriptive and exploratory research designs are the two

fundamental categories of research design. The use and adoption of

both research designs is primarily based on the nature and requirements

of the study [27].

The descriptive research design is inappropriate for this research

because of scientific necessity such as laboratory experiment. The

exploratory research design can better fit in this research because of

highlighting the links (significant or insignificant) between credit risk

and bank profitability. It is supposed that finding these links will help

the UK banks to avoid credit risks in the future. In addition, the study¡¯s

nature is elastic and distinct in answering the research questions.

Therefore, the research objectives can be accomplished more explicitly

while adopting exploratory research design [28]. Moreover, the results

of this study are largely rooted in the quantitative data and thus

careful and thorough investigation is needed which can be achieved by

adopting exploratory research.

Population

The key aim of this research is to determine the links between bank

profitability and credit risks associated with banks. The numerical data

for analyses is acquired from five big UK banks for the period of eight

years starting from 2007 to 2015. The big five UK banks refer top five

UK commercial banks which include:

1. HSBC

2. Barclays

3. Royal Bank of Scotland (RBS)

4. Lloyds Banking Group (consists of Lloyds TSB, Halifax, and

Bank of Scotland)

5. Standard Chartered Bank (SCB)

The updated list of top UK banks is acquired from the MarketCensus.

com. The data required for selected variables is acquired from Bank

Scope database which extracts information from the financial

statements of the banks. The financial statements of some banks were

Volume 5 ? Issue 2 ? 1000192

Citation: Saeed MS, Zahid N (2016) The Impact of Credit Risk on Profitability of the Commercial Banks. J Bus Fin Aff 5: 192. doi:10.4172/21670234.1000192

Page 4 of 7

also considered to find double check the information/data extracted.

Data collection

The empirical data about study variables for the period of eight

years (2007 to 2015) is collected from Bank Scope database which

contains the data of all commercial UK banks. This period is important

because it covers the financial crises of 2008 as well. Two types of

empirical data (dependent variables and independent variables) are

collected based on the theoretical model of the study (Figure 1). The

dependents variables are ROE (Return on Equity) and ROE (Return on

Assets) and conversely the independent variables are the factors that

affect bank profitability including the credit risk. So the independent

variables include credit risk variables, bank size, growth, and leverage

as shown in Table 1.

Model

This study investigates the impact of credit risk on profitability of

big five UK commercial banks. For this purpose, it is essentially required

to find the relationship between credit risk and profitability indicators

and that is why the regression model is used to declare dependent and

independent factors. A general linear model of regression is outlined in

equation 1 where ¡®Y¡¯ indicates the dependent variables and ¡®X¡¯ are the

independent factors. ¡®C¡¯ shows the coefficient.

Y=c+f(X)

(1)

By putting the study variables in above equation, two equations 2

and 3 can be formed where ROEi,t and ROAi,t represent the profitability

factors (i=1,¡­N, and t=1,¡­,T) which depend upon the independent

factors such as credit risk factors including CRIMP (Credit risk

calculated as impairments divided by total loans), CRNPL (Credit risk

calculated as non-performing loans divided by total assets), BS (bank

size), GR (growth in bank interest income), and LV (leverage ratio).

ROAi,t=c0+¦ÁCRIMP i,t+¦ÂCRNPL i,t+¦ÖBS i,t+¦ÄGR i,t+¦ÃLVi,t+ i,t

(2)

ROEi,t=c0+¦ÁCRIMP i,t+¦ÂCRNPL i,t+¦ÖBS i,t+GR i,t+¦ÃLVi,t+ ¦Åi,t

(3)

Apart from the regression analysis, the descriptive and correlation

analyses are also performed. The descriptive analyses indicate the

calculation of fundamental statistical formulas such as central

tendencies like mean, median, mode; and deviations like standard

deviation. The central tendencies show the averages of the particular

variables while standard deviation indicates the variability of data or

the standard error.

The links or associations between credit risk variables and

profitability indicators can be found through correlation analysis. The

correlation analysis in this study is used to find the association of each

profitability indicator (i.e. ROA and ROE) with all credit risk variables.

The formula of correlation is as follows which is given by Karl Pearson.

Independent

variables

Dependent

variables

r=

¡Æ ( x ? x)( y ? y)

¡Æ ( x ? x) ? ¡Æ ( y ? y )

2

2

(4)

The Karl Pearson¡¯s formula of coefficient of correlation (r) is

popular for finding correlations and according to its assumption the

results should remain within the range of -1 to +1. The results near to

+1 show stronger correlation or links between variables, and results

close to -1 point out weaker relationships. Moreover, if the result of ¡®r¡¯

is perfectly zero then it shows that both variables have no relationships

at all (Peck et al. 2011).

Results and Discussion

Regression results

The regression model considered two profitability measures ROA

and ROE which depend upon 5 independent credit risk indicators

including: CRIMP ¨C Credit Risk due to net off-charge or impairments,

CRNPL ¨C Credit risk due to non-performing loans, BS ¨C bank size, GR

¨C growth and LV ¨C leverage. Table 2 indicates independent variables as

credit risk indicators which are entered into both regression equations

(i) and (ii) (Table 2).

ROAi,t=c0+¦ÁCRIMP i,t+¦ÂCRNPL i,t+¦ÖBS i,t+¦ÄGR i,t+LVi,t+ ¦Åi,t

(i)

ROEi,t=c0+¦ÁCRIMP i,t+¦ÂCRNPL i,t+¦ÖBS i,t+GR i,t+¦ÃLVi,t+ ¦Åi,t

(ii)

Two regression models are indicated in Table 3 showing the

variability percentage of independent variables. The ¡°R square¡±

demonstrates the relationship between dependent and independent

variables whereas ¡°R¡± represents the square root of R. The value of R

points out how independent variables are associated to ROA and ROE.

Moreover, the ¡°adjusted R square¡± mentions the statistical shrinkage of

credit risk variables. Simply, adjusted R square refers the compatibility

of independent variables with dependent ones in order to validate the

decisions based on regression model [29].

In model 1, the value of R square is 0.281 which demonstrates a

suitable level of association between all the variables. The shrinkage

level for model 1 is 0.089 (8.9%) which is calculated by taking difference

of R square and adjusted R square values. In fact, there is no hard and

fast rule for assessing the shrinkage level; however, it is acceptable if lies

between 10 and 15% [30]. The shrinkage level of model 1 is between

this specific range and therefore accepted because it represents the

significance of variables involved as predictors (Table 3).

The value of R square in model 2 is 0.154 and the shrinkage level is

0.106 (10.6%) which is relatively higher then model 1 but lies between

10 and 15% and thus accepted. Both models are similar in terms of R,

R square and adjusted R square but standard error of the estimate of

model 2 demonstrates high value as compared to model 1. This shows

the significance of the effect of random changes [31].

The ANOVA analysis in Table 4 shows the statistical significance

of predictors (or independent factors) and their unpredictability over

ROA and ROE. This significance is showed in Table 4 using ¡®F¡¯ and

¡°Sig.¡± values. The ¡°Sig.¡± value is also known as P-Value. In model 1, the

p-value 0.007 is below 0.01 and 0.05 standards which shows that the

Measure

Formula

ROA

=Net income/Total assets

Impact

Source

Bank scope

ROE

=Net income/Shareholder equity

Bank scope

Credit risk

=Net Charge Off (impairments)/Total loans and advances (to customers and banks)

+/-

Bank scope

Credit risk

=Non-performing loans/Total loans and advances (to customers and banks)

+/-

Bank scope

Bank size

=Total assets

+

Bank scope

Growth

=Growth in net interest income of bank

+

Bank scope

Leverage

=Total debt/Total assets

+

Bank scope

Table 1: Study variables.

J Bus Fin Aff

ISSN: 2167-0234 BSFA an open access journal

Volume 5 ? Issue 2 ? 1000192

Citation: Saeed MS, Zahid N (2016) The Impact of Credit Risk on Profitability of the Commercial Banks. J Bus Fin Aff 5: 192. doi:10.4172/21670234.1000192

Page 5 of 7

Model

Variables Entered

(i) and (ii)

CRIMP, CRNPL, BS, GR, LVa

Variables Removed

Method

Enter

Table 2: Variables entered/removed.

Model

R

R Square

Adjusted R Square

Std. Error of the

Estimate

1

0.530a

0.281

0.192

4.01

2

0.391a

0.154

0.048

16.801

Correlation analysis

a. Predictors: (Constant), CRIMP, CRNPL, BS, GR, LV.

Table 3: Summary of the models.

Model

1

2

Sum of Squares

Mean Square

Regression

388.62

50.822

Residual

1014.054

16.578

Total

1398.674

Regression

3115.714

392.937

Residual

17880.561

281.423

Total

20984.734

F

Sig.

3.202

0.007a

1.419

0.222a

Predictors: (Constant), CRIMP, CRNPL, BS, GR, LV.

Dependent Variable: ROA.

c

Dependent Variable: ROE.

a

b

Table 4: ANOVAb,c.

Variables

Constant

ROA

ROE

Coefficients

t-value

Sig.

Coefficients

t-value

Sig.

76.514

1.690

0.095

109.811

0.572

0.573

CRIMP

0.069

1.27

0.894

0.715

1.202

0.247

CRNPL

0.210

0.370

0.722

0.713

-1.170

0.248

BS

0.059

0.521

0.610

0.138

1.050

0.300

GR

0.178

1.362

0.175

0.289

2.110

0.039

LV

0.340

1.491

0.134

0.134

0.570

0.633

Table 5: Coefficients.

relationship between predictor variables and dependent factors. The

F-value 3.202 in Table 4 denotes an appropriate link between dependent

and independent factors in model 1. However, model 2 demonstrates

0.222 p-value which is higher than 0.01 and 0.05 standards. This means

that the association between dependent and independent variables

is non-linear. In contrast, the F-value 1.419 indicates apt level of

association between variables (Table 4).

Table 5 provides detail of beta coefficients of model 1 and model

2 of regression. Based on Table 5 coefficients, the following regression

models of ROA and ROE are formed.

ROA=76.514+0.069 (CRIMP)+0.210 (CRNPL)+0.059 (BS)+0.178

(GR)+0.340 (LV) (Model 1)

ROE=109.811+0.715 (CRIMP)+0.713 (CRNPL)+0.138 (BS)+0.289

(GR)+0.134 (LV) (Model 2)

In both models, all credit risk variables have positive impact

on ROA and ROE. These results are similar to various researchers

including Sinkey and Greenwalt, Ahmed et al, Berr¨ªos and Ueda and

Mauro (Table 5) [14,32-34].

Validity of regression results

Multicollinearity statistics is the reliable measure to calculate the

validity of regression analysis and this is usually done through SPSS.

Using multicollinearity statistics, the Variance Inflation Factor (VIF)

and tolerance level are calculated (Table 6). The outcomes in Table 6

can be evaluated on the particular criteria. For instance, it is suggested

J Bus Fin Aff

ISSN: 2167-0234 BSFA an open access journal

by Gujarati that the VIF value should be under 5 and the 1/VIF (or

multicollinearity) value should be nearer to zero [35]. If these conditions

are met then the regression analysis is considered to be validated. As

shown in Table 6 that VIFs of variables is under 5 and 1/VIF values

are also nearer to zero. This shows the absence of multicollinearity in

regression analysis (Table 6).

The correlation analysis is done to correlate ROE and ROA

profitability indicators with credit risk factors that are considered

independent in this research. Therefore, this section is divided into two

subsections:

? Correlating ROA with credit risk factors

? Correlating ROE with credit risk factors

The correlation matrix in Table 7 carries the correlation between

ROA and credit risk variables having influence on bank profitability.

Table 7 shows that all factors including Net off-charge impairments,

non-performing loans, bank size, growth and leverage are positively

correlated with ROA. However, bank size and leverage have weak

association as compared to other factors. It is evident in Table 7 that

all other factors are also positively correlated with each other apart

from leverage and net-off charge impairments. These two have slightly

negative relationship which is not a big issue. These correlation results

are similar to various studies conducted in the past where Sinkey and

Greenwalt and Berr¨ªos are prominent (Table 7) [14,33].

Like ROA, ROE is another measure to quantify profitability. The

correlation matrix in Table 8 demonstrates the correlation between

ROE and credit risk variables. It is shown in the Table 8 that all

credit risk factors (apart from Net off-charge impairments ¨C CRIMP)

are positively correlated with ROE. However, they have no strong

Variables

Variance Inflation Factor

(VIF)

1/VIF

CRIMP

4.848

0.206271

CRNPL

3.752

0.266525

BS

1.822

0.248847

GR

1.349

0.34129

LV

4.640

0.215517

Table 6: Multicollinearity Statistics.

ROA

ROA

CRIMP

CRNPL

BS

GR

LV

1

CRIMP

0.223

1

CRNPL

0.211

0.98

1

BS

0.015

0.036

0.054

1

GR

0.429

0.095

0.116

0.171

1

LV

0.082

-0.028

-0.044

0.77

0.558

1

Banks=5, Variables=6

Table 7: ROA relationships.

ROA

ROA

CRIMP

CRNPL

BS

GR

LV

1

CRIMP

-0.006

1

CRNPL

0.041

0.704

1

BS

0.155

0.043

0.126

1

GR

0.176

0.048

0.137

0.271

1

LV

0.087

-0.085

-0.019

0.782

0.45

1

Banks=5, Variables=6

Table 8: ROE relationships.

Volume 5 ? Issue 2 ? 1000192

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download